INDEX

Explanations

phrases related to specific styles or themes, including mentions of particular characters or concepts

oai_token-act-pair · gpt-3.5-turbo

negative connotations or themes

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Rica

-0.74

 Indigo

-0.74

 Discipline

-0.73

 Fiscal

-0.71

 Confederation

-0.71

 Dickinson

-0.69

 Levi

-0.68

 Moody

-0.68

sd

-0.67

Ud

-0.66

POSITIVE LOGITS

themed

1.33

style

1.28

based

1.27

inspired

1.25

centric

1.23

loving

1.19

related

1.17

hun

1.17

sized

1.16

esque

1.15

Activations Density 0.182%

phrases related to specific styles or themes, including mentions of particular characters or concepts

negative connotations or themes

No Comments

No Known Activations